Matching Internal Page Links

[ocon]

I am writing a script to identify and perform an action on internal page links. I've written my script below but I'm not comfortable with how it identifies internal links.

for(i = 0; i < document.links.length; i++){
 if(document.links[i].href.indexOf(window.location.host+window.location.pathname+'#') !== -1){...}}

I need it to match the following urls:

#
#anchor
http://www.mysite.tld/file1.html
http://www.mysite.tld/file1.html#
http://www.mysite.tld/file1.html#anchor

But not:

file2.html
http://www.mysite.tld/file2.html
http://www.othersite.tld/file3.html
file2.html#anchor
http://www.mysite.tld/file2.html#anchor
http://www.othersite.tld/file3.html#anchor

Is there a better way to match internal links than:

document.links[i].href.indexOf(window.location.host+window.location.pathname+'#') !== -1

[daveVk]

But not:
file2.html
h t t p://www.mysite.tld/file2.html
h t t p://www.othersite.tld/file3.html
file2.html#anchor
h t t p://www.mysite.tld/file2.html#anchor
h t t p://www.othersite.tld/file3.html#anchor

What aspect of red entries do you consider makes them NOT "internal page links", I to do see pattern ?

[phranque]

What aspect of red entries do you consider makes them NOT "internal page links"

i think ocon means on-page links - as in links to the url of the page itself or links to fragment identifiers on the current page.

[daveVk]

Sorry somehow I read it as same site links.

What you have looks fine but will not match links in the trivial case of no '#',

assuming you are on page

http://www.mysite.tld/file1.html

will not match

http://www.mysite.tld/file1.html
http://www.mysite.tld/file1.html?id=7

etc

If site uses query strings, the query part of the the URL needs to be added to check.

It will also catch cases like

/file1.html#
file1.html#

and other relative URL notations

[Readie]

You can regex check it - seems a bit more reliable than what you've got there (Assuming I've understood your request properly).


Disclaimer: Not tested this code, may not work out-the-box.

function regQuote() {
    return arguments[0].replace(/([\-\.])/, function(m) {
        return '\\' + m[1];
    });
}

function checkLink() {
    var href = arguments[0];
    var regexp, check;
    
    if(href.match(/^\#/)) {
        // Starts with a hash. Just an anchor on this page
        return true;
    }
    
    if(href.match(/^https?:\/\//)) {
        // If it starts with http(s)://, we need to check the domain
        regexp = new RegExp('^https?:\\/\\/' + regQuote(window.location.host) + '($|/|\\#|\\?)');
        if(!href.match(regexp)) {
            return false;
        }
    }
    
    if(href.match(/^\//)) {
        // Starts with a slash. Need to validate against entire path
        check = window.location.pathname
    } else {
        // Does not start with a slash. Just want the end segment
        check = window.location.pathname.split('/').pop();
    }
    
    regexp = new RegExp('^' + regQuote(check));
    if(!href.match(regexp)) {
        return false;
    }
    
    return true;
}

/* USAGE:

if(checkLink(document.links[i].href) === true) {
    // Link to this page
}

*/

[daveVk]

My understanding if that document.links[ i ].href will return the fully qualified URL regardless of how href is written. Original code should be reliable.

After some tests, it appears that my_a_dom_node.href always return the full-qualified URI, including the [domaine.name,...] which should be okay for what I want.

[edited by: whoisgregg at 6:45 pm (utc) on Jul 29, 2013]
[edit reason] rv link, please see sticky mail :) [/edit]

[Readie]

Ahh. I did not know that. Never mind then :)